An algorithm for automatic construction of computational procedure of non-linear local image processing on the basis of hierarchical regression
V.N. Kopenkov, V.V. Myasnikov

Full text of article: Russian language.

Abstract:
An algorithm of the automatic construction of a computational procedure of the local digital signal and image processing is presented. The computational procedure is based on a local discrete wavelet transform of the image used for the preliminary image analysis, and a hierarchical regression used to obtain the transformation results. The construction process based on the precedent usage analysis ("input" – "output" image pairs) takes into consideration the limitation on the complexity of the conversion constructed, maximizing the processing quality and generalization capability.

Key words:
local processing, hierarchical regression, computational efficiency.

References:

  1. Vapnik, V.N. Theory of Pattern Recognition / V.N. Vapnik, A.V. Chervinenkis. – Moscow: “Nauka” Publisher, 1974. – (In Russian).
  2. Vorontsov, K. A combinatorial approach to assessing the quality of training algorithms // Mathematical problems of cybernetics / Ed O.B. Lupanov. – Moskow: “Fizmatlit” Publisher. – 2004. – Vol. 13. – P. 5-36. – (In Russian).
  3. Glumov, N.I. Using of the polynomial bases for image processing in a sliding window / N.I. Glumov, V.V. My­asnikov, V.V. Sergeev // Computer Optics. – 1995. – Vol. 14-15(1). – P. 55-68. – (In Russian).
  4. Dudgeon, D.P. Multidimensional Digital Signal Processing / D.P. Dudgeon, R. Mersero. – Moscow: “Mir” Publisher, 1998. – (In Russian).
    Methods of computer image processing. Part II: Methods and algorithms, Ed. V.A. Soifer. – Moscow: “Fizmatlit” Publisher, 2009. – 784 p. – (In Russian).
  5. Pratt, W.K. Digital image processing / W.K. Pratt. – Moscow: “Mir” Publisher, 1982. – B.2. – 480 p. – (In Russian).
  6. Sergeev, V.V. Comparative analysis of methods for neural networks and hierarchical approximation in problems of image filtering / V.V. Sergeev, V.N. Kopenkov, A.V. Cher­nov // Avtometriya. – 2006. – Vol. 42, N 2. – P. 100-106. – (In Russian).
  7. Wasserman, P.D. Neural Computing. Theory and Practice / P. Wasserman. – Moscow: “Mir” Publisher, 1992. – 240 p. – (In Russian).
  8. Gonzalez, R. Digital image processing / R. Gonzalez and R. Woods. – Moscow: “Technosphere” Publisher, 2005. – 1072 p. – (In Russian).
  9. Ahn, H. Tree-structured exponential regression modeling / H. Ahn // Biometrical Journal. – 2007. – Vol. 36. – P. 43-61.
  10. Breiman, L. Classification and regression trees / L. Breiman, J.H. Friedman, R.A. Olshen and C.J. Stone.// Monterey, Calif., U.S.A.: Wadsworth, Inc., 1984.
  11. Haikin, S. Neural Networks: A Comprehensive Foundation / S. Haikin. – Moscow: «Vilyams», 2006. – 1104 p.
  12. Kopenkov, V. Effecient algorithms of local discret wavelet transform with HAAR-like bases / V. Kopenkov // Pattern Recognition and Image Analysis. – 2008. – Vol. 18, N 4. – P. 654-661.
  13. Kopenkov, V.N. Research the Performance of a Recursive Algorithm of the Local Discrete Wavelet Transform. / V.N. Kopenkov, V.V. Myasnikov // 20-th International Conference on Pattern Recognition (ICPR-2010). – Istambul, Turcey August 23-26, 2010. Abstract book. – P. 317.
  14. Kopenkov, V.N. Regression restoration methods as applied to solve the problem of multidimensional indirect measurements / V.N. Kopenkov, V.V. Sergeev, E.I. Tim­bai // Pattern Recognition and Image Analysis. – 2011. – Vol. 21, N 2. – P. 480-483.
© 2009, IPSI RAS
Institution of Russian Academy of Sciences, Image Processing Systems Institute of RAS, Russia, 443001, Samara, Molodogvardeyskaya Street 151; E-mail: ko@smr.ru; Phones: +7 (846 2) 332-56-22, Fax: +7 (846 2) 332-56-20